Skip to main content
NCBI home page
Elsevier - PMC COVID-19 Collection logoLink to Elsevier - PMC COVID-19 Collection
. 2021 Dec 2;915:174670. doi: 10.1016/j.ejphar.2021.174670

In vitro ion channel profile and ex vivo cardiac electrophysiology properties of the R(-) and S(+) enantiomers of hydroxychloroquine

Véronique Ballet a, G Andrees Bohme b,∗, Eric Brohan c, Rachid Boukaiba b, Jean-Marie Chambard b, Odile Angouillant-Boniface c, Thierry Carriot a, Céline Chantoiseau b, Sophie Fouconnier b, Sylvie Houtmann b, Céline Prévost c, Brigitte Schombert b, Laurent Schio d, Michel Partiseti b
PMCID: PMC9749463  PMID: 34863995

Abstract

Hydroxychloroquine (HCQ) is a derivative of the antimalaria drug chloroquine primarily prescribed for autoimmune diseases. Recent attempts to repurpose HCQ in the treatment of corona virus disease 2019 has raised concerns because of its propensity to prolong the QT-segment on the electrocardiogram, an effect associated with increased pro-arrhythmic risk. Since chirality can affect drug pharmacological properties, we have evaluated the functional effects of the R(-) and S(+) enantiomers of HCQ on six ion channels contributing to the cardiac action potential and on electrophysiological parameters of isolated Purkinje fibers. We found that R(-)HCQ and S(+)HCQ block human Kir2.1 and hERG potassium channels in the 1 μM–100 μM range with a 2–4 fold enantiomeric separation. NaV1.5 sodium currents and CaV1.2 calcium currents, as well as KV4.3 and KV7.1 potassium currents remained unaffected at up to 90 μM. In rabbit Purkinje fibers, R(-)HCQ prominently depolarized the membrane resting potential, inducing autogenic activity at 10 μM and 30 μM, while S(+)HCQ primarily increased the action potential duration, inducing occasional early afterdepolarization at these concentrations. These data suggest that both enantiomers of HCQ can alter cardiac tissue electrophysiology at concentrations above their plasmatic levels at therapeutic doses, and that chirality does not substantially influence their arrhythmogenic potential in vitro.

Keywords: CiPA initiative, Torsade de pointes, Patch-clamp, Purkinje fibers, Covid-19, Chiral switch

1. Introduction

Hydroxychloroquine (HCQ, Plaquenil®) is a well-known antimalarial drug currently used in immuno-inflammatory conditions such as cutaneous and systemic lupus erythematosus or rheumatoid arthritis (Schrezenmeier and Dorner, 2020). Recent attempts to repurpose HCQ for the treatment of corona virus disease 2019 (Covid-19) have raised concerns because this drug can prolong the QT-segment on the electrocardiogram (Mazzanti et al., 2020; Oscanoa et al., 2020). Drug-induced QT interval prolongation is associated with an increased risk of a distinctive type of ventricular arrythmias known as Torsade de Pointes (TdP) which can lead to cardiac arrest (Roden, 2004). The mechanism underlying TdP results from a delay of the cardiac action potential repolarization towards the end of its plateau phase, which prematurely reactivates the slowly inactivating L-type CaV1.2 calcium channels. Action potential repolarization is driven by outward potassium currents flowing through channels encoded by the human ether-a-go-go related gene (i.e. hERG channels). Inhibition of hERG function was the first molecular mechanism associated with TdP Rampe and Brown (2013); (Sanguinetti et al., 1995). However, drug effects at the voltage gated KV7.1 or the inwardly rectifying Kir2.1 channels can also influence repolarization and induce arrhythmic events. Moreover, blockade of inward Na+ and Ca2+ currents could in principle counterbalance the QT elongating effects of outward K+ currents inhibition, a hypothesis that underlies the Comprehensive in vitro Pro-Arrhythmia (CiPA) initiative aimed at refining the electrocardiographic risk assessment of new drugs (Gintant et al., 2016).

HCQ is a chiral 4-aminoquinoline which has been developed and marketed as a racemate. Yet, it has been documented for several racemic drugs that the pharmacology of each enantiomer can give rise to a eutomer and a distomer where the desired properties can be ascribed primarily to the Rectus- or the Sinister-enantiomer (Tucker, 2000). Chirality can also influence pharmacodynamic behavior, a feature that has been demonstrated for R(-)HCQ and S(+)HCQ in the mid-90's (Ducharme et al., 1994; Tett et al., 1994). Several methods have been described to obtain the enantiomers of HCQ separately, including direct synthesis using optically resolved reactants (Blaney et al., 1994) and chromatographic separation on a preparative scale (Wilson et al., 2020).

Here we have evaluated the effects of R(-)HCQ and S(+)HCQ in in vitro electrophysiological models pertaining to cardiovascular safety pharmacology. First, we evaluated the effects of each enantiomer on six cardiac ion channels using automated patch-clamp and cloned human channels expressed in mammalian cells. Secondly, we tested the effects of R(-)HCQ and S(+)HCQ on transmembrane currents recorded from ex vivo isolated rabbit Purkinje fibers, a native tissue that recapitulates the essential features of human ventricular conductive tissue (Baczko et al., 2016). We found that the enantiomers of HCQ selectively inhibit Kir2.1 channels and, less potently, hERG channels, and significantly alter several cardinal AP metrics at concentrations about one order of magnitude above the free plasma concentrations circulating at therapeutically effective doses.

2. Material and methods

2.1. Chiral preparative HPLC

Separation of the enantiomers of HCQ was performed using a carbamate-derivatized amylose chiral stationary phase (Chiralpak® AD, Daicel chemical industries). About 1 g hydroxychloroquine sulfate was dissolved in 200 mL of 90:10 acetonitrile/methanol containing 5% triethylamine and loaded in two halves on a 76.5 mm × 350 mm column packed with 1.2 kg chiral phase at 20 μm particle size. Elution was performed with the same solvent mix containing 0.1% triethylamine at a flowrate of 400 mL/min. Efficiency was improved by recycling the feed after the desorption of the first enantiomer. This second passage on the same column allowed complete resolution of both enantiomers. The fractions monitored at 265 nm UV were collected and evaporated to obtain free-base oils. Absolute configurations were assigned as S(+)HCQ (optical rotation OR = +76.4 [c] = 320 g/mL in DMSO at 589 nm; enantiomeric excess ee = 98.5%) and R(-)HCQ (OR = −68.1 same conditions; ee = 99%) based on published data (Ibrahim and Fell, 1990).

2.2. Patch-clamp on human cardiac ion channels

Studies were performed as described previously (Le Marois et al., 2020). Briefly, CHO cells expressing hERG or KV7.1/minK were obtained from B'Sys (Switzerland), and CaV1.2/β2/α2δ1, Kir2.1 or KV4.3/KChIP2.2 from Charles River (USA), respectively. HEK cells expressing NaV1.5 were constructed in house. All expressions were constitutive except Kir2.1, CaV1.2 and NaV1.5 which were induced overnight by exposure to 1 μg/mL doxycycline (in the presence of 3 μM verapamil for CaV1.2). Whole-cell recordings were performed at room temperature on a Qpatch® 48X planar patch-clamp workstation (Sophion, Denmark). The pulse-protocols used and cycle-times applied to elicit each current are schematized as insets in their respective figures. Leak currents were handled either on-line by P/n subtraction or offline by subtracting currents remaining in the presence of a specific reference inhibitor added at the end of each recording session at a maximally active (i.e. “full-block”) concentration. Extracellular buffers for potassium channels contained (in mM) NaCl, 150; KCl, 4; CaCl2, 2; MgCl2, 1 and HEPES, 10. For the hERG channels, glucose (10) was added. The pH was adjusted to 7.4 with NaOH. The intracellular buffers were composed of (in mM): KF, 120; KCl, 20; EGTA, 10; MgCl2, 1; HEPES, 10. For KV7.1 channels, 10 mM EDTA was added as a second chelator, EGTA was decreased to 5 mM and no MgCl2 was added. The pH was adjusted to 7.2 with KOH. The extracellular buffer for CaV1.2 channels contained (in mM): NaCl, 145; KCl, 4; BaCl2, 10 and HEPES, 10, and pH was adjusted to 7.4 with NaOH. The intracellular buffer contained: CsF, 27; CsCl, 112; EGTA, 8.2; NaCl, 2; HEPES, 10 and Mg-ATP, 4. The pH was adjusted to 7.2 with CsOH. For NaV1.5 channels, the extracellular buffer contained: NaCl, 137; KCl, 4; CaCl2, 2; MgCl2, 1 and HEPES, 10. The pH was adjusted to 7.4 with NaOH. The intracellular buffer contained: CsF, 150; EGTA/CsOH, 1/5; NaCl, 10; MgCl2, 1; CaCl2, 1; HEPES, 10. The pH was adjusted to 7.2 with CsOH. All channels were exposed to six concentrations of R(-)HCQ or S(+)HCQ applied cumulatively in an ascending order up to 90 μM. Specific inhibitors applied at full-block concentrations to establish assay sensitivity and isolate leak currents were E−4031 (10 μM) for the hERG channels, lidocaine (3 mM) for the NaV1.5 channels, CdCl2 (0.2 mM) for CaV1.2, HMR-1556 (30 μM) for KV7.1, SKF-96365 (30 μM) for KV4.3 and BaCl2 (3 mM) for the Kir2.1 channels. Bepridil was used as a positive control and tested at concentrations up to 30 μM in parallel to the HCQ enantiomers for all channels which were insensitive to them. Treatments were done in buffer containing 0.3% or 1% DMSO, and 1% Pluronic F-68. All inhibitions were quantified as change in normalized peak current amplitude except for KV4.3 currents which were quantified as change in normalized integral charge transferred calculated as the area under the current trace versus time.

2.3. Current clamp on excised rabbit Purkinje fibers

Methods were reviewed and endorsed by Sanofi's internal animal ethics committee and conducted in accordance with European Directives 86/609/EEC and 2010/63/EU under specific approval by the French Ministry of Research (APAFiS authorization N° 2018080110555740). As previously described (Le Marois et al., 2020), Purkinje fibers were dissected out from excised hearts of male New Zealand rabbits weighing 1.7–2.1 kg and fixed in a small ex vivo tissue bath continuously perfused with a buffer containing (in mM): NaCl 120; KCl 4; MgCl2 1; NaH2PO4 1.8; NaHCO3 25; glucose 11; CaCl2 1.8 maintained at 36 ± 1 °C and delivered at a flow-rate of 10 mL/min. Oxygen supply and physiological pH 7.4 were obtained by bubbling 95% CO2/5% O2 in the buffer reservoirs. The tissue action potential (AP) firing was driven by electrical pulses delivered through stainless-steel electrodes. The membrane resting potential, AP amplitude, depolarization speed (Vmax), and AP duration at 50% and 90% repolarization (ADP50 and APD90, respectively) were recorded using 3 M KCl-filled fine glass micro-electrodes and standard signal amplification and computerized acquisition and monitoring devices. Concentrated (12 mM) stock solutions of S(+)HCQ or R(-)HCQ were prepared in 100% DMSO before being formulated in study buffer to the final concentrations tested. Each fiber (n = 6 per drug) was exposed to four increasing concentrations of R(-)HCQ or S(+)HCQ according to the protocol depicted in Fig. 7. During the recordings, 1 mL samples of the formulated buffers were collected before (in the reservoirs at room temperature) and after the fibers (at the outlet of the tissue chamber 20 min following onset of each concentration perfusion) to determine their actual drug exposure.

Fig. 7.

Fig. 7

Open in a new tab

Experimental protocol for the Purkinje fiber studies. The perfused tissue samples were first exposed to normal buffer under pacing at the indicated frequency and for the indicated durations, before being exposed to drug-containing buffer. R(-)HCQ or S(+)HCQ were tested at 0.3 μM, 3 μM, 10 μM and 30 μM, each applied for 30 min in ascending order. After the highest concentration, a 30 min washout with normal buffer was added to evaluate reversibility of drug effects. Five action potential parameters were measured at the end of each frequency period. The slowest pacing rate (i.e. 0.25 Hz or 15 beats/min) was used to favor the occurrence of abnormal repolarization events, if drug should induce any. The fastest pacing rate (i.e. 3 Hz or 180 bpm) was used to evaluate use-dependency of changes observed.

2.4. Data analysis

Patch-clamp current-trace captures and current-size measurements were exported from Sophion's proprietary Assay Software for Qpatch to a professional graphing package (Prism 8.3.0, GraphPad Software, San Diego, CA, USA) which served also for half-maximal inhibitory concentration (IC50) estimations by fitting percent inhibition data (normalized to pre-drug levels) to a sigmoidal curve with minimum and maximum constrained to 0% and 100%, respectively. Calculated IC50 values are reported along with their 95% confidence interval (95% CI). Statistical analyses of the Purkinje fiber data were performed by one-way repeated factor analysis of variance (ANOVA) followed by Dunnett's test versus baseline for each drug at each pacing rate using SAS 9.4 software (SAS Institute Inc., Cary, NC, USA). A rank-transformation was applied in case of normality hypothesis violation or lack of variance homogeneity. Due to a high number of missing values, the 30 μM data set for R(-)HCQ was excluded from statistical analysis at all pacing rates and for all metrics except for the resting potential data. Likewise, the 10 μM data set for R(-)HCQ at the lowest pacing rate was analyzed only for the resting potential due to missing data for the other parameters. Significance levels were set to 5%. Descriptive statistics and graphs were generated with Prism 8.3.0.

3. Results

3.1. Ion channel studies

Cloned human cardiac ion channels expressed in CHO or HEK cells were recorded on a QPatch® 48X automated patch-clamp workstation in population patch mode. Fig. 1 A illustrates the effects of R(-)HCQ and S(+)HCQ on hERG current traces elicited by a classical two-step long-duration (5 s per step) voltage-protocol allowing ample time for channel blockade at room temperature and triggering large outward tail currents upon its + 20 mV to −50 mV repolarizing step. Both R(-)HCQ and S(+)HCQ concentration-dependently inhibited these currents. The time-course of the effect of each of the six concentrations applied cumulatively indicated that the active concentrations could reach steady-state inhibition at the end of each exposure period (Fig. 1B). The final application of the selective and potent hERG blocker E−4031 (1 μM) eliminated all currents, thereby establishing the sensitivity of the assay. The half-maximal inhibitory concentrations (IC50) estimated from the sigmoidal curve fitting of the percent block data indicate a 3.7-fold greater potency of R(-)HCQ over S(+)HCQ (Table 1 ).

Fig. 1.

Fig. 1

Open in a new tab

R(-)HCQ and S(+)HCQ inhibit hERG potassium currents. Panel A: representative outward currents generated by the voltage-command protocol schematized at reduced scale above the traces in the absence (vehicle, green current trace) or presence of 30 μM of either enantiomer (blue and red traces for R(-)HCQ and S(+)HCQ, respectively). Panel B: time-course of hERG tail current amplitude in response to cumulative application of six increasing concentrations of R(-)HCQ or S(+)HCQ, preceded by three applications of vehicle (Veh.) to test current stability, and followed by a final application of a reference hERG inhibitor (E−4031) at full-block concentration (10 μM). Current traces and symbol colors are matched in panels A and B.

Table 1.

Inhibitory potencies of R(-)HCQ and S(+)HCQ on six cardiac ion channels of the Comprehensive in vitro Pro-arrhythmia Assay (CiPA) panel.

Channel R(-)HCQ N = S(+)HCQ N =
hERG 20 μM [19–22] 7 74 μM [65–86] 5
NaV1.5 >90 μM 6 >90 μM 7
CaV1.2/β2/α2δ1 >90 μM 6 >90 μM 5
KV4.3/KChIP2.2 >90 μM 7 >90 μM 7
KV7.1/minK >90 μM 4 >90 μM 5
Kir2.1 3.3 μM [2.9–3.7] 10 6.5 μM [5.7–7.6] 6
Open in a new tab

Subunit compositions of the human recombinant channels tested are detailed in first column. IC50 values are shown along with their 95% confidence intervals.

Fig. 2 summarizes the effects of the enantiomers on human NaV1.5 channels expressed in HEK cells. These were held at −80 mV outside of the depolarizing voltage-steps to populate both the closed and inactivated states of the channels as described (Donovan et al., 2011). The fast inactivating inward currents were generated by short depolarizing steps to −40 mV. Interestingly, neither R(-)HCQ nor S(+)HCQ affected NaV1.5 to a meaningful extent (i.e. below about 10% decrease which can be attributed to current rundown) up to the highest tested concentration of 90 μM. In contrast, the local anesthetic and NaV1.5 blocker lidocaine at 3 mM systematically abolished all currents in each recording, confirming the sensitivity of the experimental conditions employed. Furthermore, we evaluated in parallel the multi-channel inhibitor bepridil, once used as an antianginal agent, and found an IC50 value of 2.5 μM along with a 95% confidence interval ranging from 2.3 μM to 2.7 μM, which is fully in accordance with published data (Crumb et al., 2016).

Fig. 2.

Fig. 2

Open in a new tab

R(-)HCQ and S(+)HCQ do not alter NaV1.5 sodium currents. Panel A: representative inward currents elicited by the voltage-command protocol schematized at reduced scale above the traces in the presence of the indicated concentrations of enantiomers (blue and red traces for R(-)HCQ and S(+)HCQ, respectively) or of 10 μM of the NaV1.5 inhibitor bepridil (black trace). Panel B: time-course of peak NaV1.5 current amplitude in response to cumulative application of six increasing concentrations of R(-)HCQ or S(+)HCQ, preceded by three applications of vehicle (Veh.) to test current stability, and followed by a final application of a reference NaV1.5 inhibitor (lidocaine) at full-block concentration (3 mM). Current traces and symbol colors are matched in panels A and B. Note that a half-log left-shifted concentration-range was tested for the positive control bepridil compared to the concentration-range tested for the enantiomers (see arrow labels in graph).

We then went on to examine the effects of HCQ's enantiomers on CaV1.2 channels expressed in CHO cells along with their β2 and α2δ1 regulatory proteins. During the patch-clamp recordings, the cells were maintained at a depolarized holding of - 50 mV to enrich the population of channels in the inactivated state as described (Kuryshev et al., 2014). The slowly inactivating inward currents were obtained by stepping the membrane to 0 mV with 200 ms square pulses. To minimize current rundown over time, barium was used as a charge carrier. Fig. 3 A depicts the current traces obtained in the presence of 90 μM R(-)HCQ or 90 μM S(+)HCQ. Neither of the enantiomers was able to block CaV1.2 currents, while CdCl2 (0.2 mM) applied at the end of the recording sessions consistently inhibited all the currents (Fig. 3B). Moreover, the positive control bepridil tested in parallel with the enantiomers was active, exhibiting an IC50 value of 5.2 μM [95% CI: 4.4 μM–6.2 μM] in agreement with published values (Crumb et al., 2016).

Fig. 3.

Fig. 3

Open in a new tab

R(-)HCQ and S(+)HCQ do not alter CaV1.2 calcium currents. Panel A: representative inward currents elicited by the voltage-command protocol schematized at reduced scale above the traces in the presence of the indicated concentrations of enantiomers (blue and red traces for R(-)HCQ and S(+)HCQ, respectively) or of the CaV1.2 inhibitor bepridil (black trace). Panel B: time-course of maximal CaV1.2 current amplitude in response to cumulative application of six increasing concentrations of R(-)HCQ or S(+)HCQ, preceded by three applications of vehicle (Veh.), and followed by a final application of CdCl2 at full-block concentration (0.2 mM). Current traces and symbol colors are matched in panels A and B. Bepridil was applied at a left-shifted concentration-range compared to the enantiomers (see arrow labels in graph).

We then profiled R(-)HCQ and S(+)HCQ on KV4.3 channels. CHO cells expressing these channels along with KChIP2.2 accessory subunits were held at −80 mV and stepped to +20 mV using 500 ms long square pulses. The latter generated large, rapidly inactivating outward currents as expected (Fig. 4 A). Unlike bepridil, neither 90 μM R(-)HCQ nor 90 μM S(+)HCQ were able to affect the current peak amplitude nor inactivation. Fig. 4B depicts the change of the total charge transferred (i.e. the integral of the area-under- the-current versus time curve) as a function of applications of each drug concentration and ultimately of 30 μM SKF-96365, a non-selective cationic channel blocker with micromolar potencies at several cardiac potassium channels including human KV4.3 (Liu et al., 2016). Bepridil, again tested in parallel, exhibited a concentration-dependent inhibitory effect on KV4.3 currents, yielding an IC50 value of 11 μM [95% CI: 9.7 μM–12 μM] (Fig. 4B).

Fig. 4.

Fig. 4

Open in a new tab

R(-)HCQ and S(+)HCQ do not alter KV4.3 potassium currents. Panel A: representative outward currents elicited by the voltage-command protocol schematized at reduced scale above the traces in the presence of the indicated concentrations of enantiomers (blue and red traces for R(-)HCQ and S(+)HCQ, respectively) or of the KV4.3 inhibitor bepridil (black trace). Panel B: time-course of change in area under the current trace (i.e. total charge transferred) in response to cumulative application of six increasing concentrations of R(-)HCQ or S(+)HCQ, preceded by three applications of vehicle (Veh.) to stabilize current, and followed by a final application of SKF-96365 (SKF) at full-block concentration (30 μM). Current traces and symbol colors are matched in panels A and B. Bepridil was applied at a left-shifted concentration-range compared to the enantiomers (see arrow labels in graph).

To further explore the activities of R(-)HCQ and S(+)HCQ on repolarizing currents of the cardiac AP, we patched CHO cells expressing KV7.1 subunits along with the ancillary protein minK which is needed to reproduce the native I Ks current (Fig. 5 A). The cells were stepped to +40 mV from a holding potential of −80 mV using prolonged 2 s square-pulses. The latter triggered a slowly activating current that resisted the application of either 90 μM R(-)HCQ or the same concentration of S(+)HCQ, while the positive control bepridil was active as expected, giving an IC50 value of 7.2 μM [95% CI: 6.6 μM–7.8 μM] (Fig. 5B).

Fig. 5.

Fig. 5

Open in a new tab

R(-)HCQ and S(+)HCQ do not alter KV7.1 potassium currents. Panel A: representative outward currents elicited by the voltage-command protocol schematized at reduced scale above the traces in the presence of the indicated concentrations of enantiomers (blue and red traces for R(-)HCQ and S(+)HCQ, respectively) or of the KV7.1 inhibitor bepridil. Panel B: time-course of change in normalized outward current amplitude measured at the end of a 2s-long depolarizing pulse in response to cumulative application of six increasing concentrations of R(-)HCQ or S(+)HCQ preceded by three applications of vehicle (Veh.) to stabilize current, and followed by a final application of HMR-1556 (HMR) at full-block concentration (30 μM). Current traces and symbol colors are matched in panels A and B with blue traces and circles for R(-)HCQ, red traces and squares for S(+)HCQ, and black traces and triangles for bepridil which was applied at a left-shifted concentration-range compared to the enantiomers (see arrows in graph).

Finally, we examined the effects of the enantiomers on Kir2.1 expressed in CHO cells (Fig. 6 ). The cells were maintained at a depolarized holding of −20 mV, and the outward component of the current relevant to AP repolarization was elicited by a step-ramp protocol from +40 mV down to −120 mV at −0.46 V/s followed by a short plateau. As previously noted for hERG, both R(-)HCQ and S(+)HCQ inhibited the Kir2.1-mediated repolarizing currents (Fig. 6A), although with approximately one order of magnitude stronger potency (Fig. 6B), albeit with a similarly modest enantiomeric separation (Table 1).

Fig. 6.

Fig. 6

Open in a new tab

R(-)HCQ and S(+)HCQ inhibit Kir2.1 potassium currents. Panel A: representative outward currents elicited by the hyperpolarizing ramp of the voltage-command protocol schematized at reduced scale above the traces in the absence (vehicle, green current trace) or presence of the indicated concentration of enantiomer (blue and red traces for R(-)HCQ and S(+)HCQ, respectively). Panel B: time-course of Kir2.1 peak outward current amplitude in response to cumulative application of six increasing concentrations of R(-)HCQ or S(+)HCQ preceded by three applications of vehicle (Veh.) to test current stability and followed by a final application of BaCl2 at full-block concentration (3 mM). Current traces and symbol colors are matched in panels A and B.

3.2. Purkinje fiber studies

Action potentials (AP) were recorded from rabbit Purkinje fibers with conventional KCl-buffer-filled fine glass microelectrodes manually inserted through the conductive tissue membrane. Fig. 7 schematizes the drug application sequence employed and pacing changes made to assess drug effects. Fig. 8 and Fig. 9 depict the effects of R(-)HCQ and S(+)HCQ, respectively, on five parameters analyzed at three different pacing rates, along with their statistical significance. Fig. 10 compares the percent change versus baseline observed for each of these metrics.

Fig. 8.

Fig. 8

Open in a new tab

Effects of R(-)HCQ on rabbit Purkinje fiber membrane potential and action potential (AP) parameters. Upper panels A, B and C illustrate typical AP-waveforms paced at 3 Hz (180 beats/min, bpm), 1 Hz (60 bpm) and 0.25 Hz (15 bpm), respectively, in the presence of the indicated concentrations of enantiomer. The graphs in panels D through R show median (with interquartile range) and individual values for each of five metrics (arranged vertically) at the different pacing rates (arranged horizontally) during cumulative perfusion of vehicle (black symbols), the four tested concentrations of R(-)HCQ (blue, dark yellow, red and pink symbols) and after 30 min washout with vehicle (green symbols). Symbol colors in graphs D through R match with waveform colors in panels A, B and C. Arrows in panel A, B and C highlight the strong depolarization of membrane resting potential. Typefaces on graphs denote significant differences from baseline (‡ = p < 0.05; ¶ = p < 0.01; § = p < 0.001).

Fig. 9.

Fig. 9

Open in a new tab

Effects of S(+)HCQ on rabbit Purkinje fiber membrane potential and action potential (AP) parameters. Upper panels A, B and C illustrate typical AP-waveforms paced at 3 Hz (180 beats/min, bpm), 1 Hz (60 bpm) and 0.25 Hz (15 bpm), respectively, in the presence of the indicated concentrations of enantiomer. The graphs in panels D through R show median (with interquartile range) and individual values for each of five metrics (arranged horizontally) at the different pacing rates (arranged vertically) during cumulative perfusion of vehicle (black symbols), the four tested concentrations of S(+)HCQ (blue, dark yellow, red and pink symbols) and after 30 min washout with vehicle (green symbols). Symbol colors in graphs D through R match with waveform colors in panels A, B and C. Arrows in panel C show occasional early after depolarizations (EADs) due to the strong action potential elongation triggering calcium channel reactivations. Typefaces on graphs denote significant differences from baseline (‡ = p < 0.05; ¶ = p < 0.01; § = p < 0.001).

Fig. 10.

Fig. 10

Open in a new tab

Effect-size of R(-)HCQ- and S(+)HCQ-induced changes in Purkinje fiber membrane and AP parameters. The graphs in panels A through O show percent change of median from baseline for each of the five metrics (arranged horizontally) measured at the different pacing rates (arranged vertically) in the presence of the four tested concentrations and after washout from R(-)HCQ (blue bars grouped on the left side of each graph) or S(+)HCQ (red bars grouped on the right side of each graph). Note the same range in vertical axis of all graphs in this figure to allow global effects comparison.

Overall, at the lowest tested concentration of 0.3 μM, neither R(-)HCQ nor S(+)HCQ had a significant effect on any of the parameters whatever the beating rate imposed (compare first and second scatter dot plots in each graph labelled D through R in Fig. 8, Fig. 9). In contrast, both enantiomers started to develop effects that occasionally reached statistical significance at 10-fold higher concentrations.

R(-)HCQ consistently decreased the resting potential of the conductive tissue membranes. This effect was concentration-dependent over the 3 μM–30 μM range, and of similar magnitude whatever the pacing rate driven by the electrical stimulations (Fig. 8, graphs D, E and F and Fig. 10, graphs A, B and C). Specifically, and for example, at the highest tested concentration of 30 μM, the median membrane potential was decreased by 23% from – 91 mV to – 70 mV at 180 beats/min (bpm), by 20% from – 90 mV to – 72 mV at 60 bpm and by 21% from – 89 mV to – 70 mV at 15 bpm. These effects were almost completely reversible after washout with vehicle buffer for most metrics (Fig. 8, compare baseline and washout scatter dot plots in graphs D, E and F). R(-)HCQ also produced a significant decrease in the AP amplitude and the depolarization velocity at 10 μM for the 180 bpm and 60 bpm pacing rates (Fig. 8, graphs G through L). However, the rather large membrane depolarization induced by R(-)HCQ resulted in the fibers escaping electrical pacing at the intermediate 60 bpm rate, and even more so at the very slow 15 bpm rate. Hence, AP amplitude, AP depolarization velocity, and APD50 and APD90 measurements could not be collected at 10 μM and 30 μM at the slowest pacing rates, and at 10 μM at the intermediate pacing rate. A nearly complete data set could only be collected from fibers stimulated at the fastest rate (180 bpm) which overrode the autogenic activity, although several data-points were still missing at 30 μM. At this rapid 180 bpm pacing rate, R(-)HCQ significantly decreased the AP amplitude and its upstroke depolarization velocity at 10 μM. The latter was halved at the highest tested concentration (– 51% at 30 μM, with n = 2 only, Fig. 10G). R(-)HCQ also elongated the AP duration by a small, but significant extent at near complete repolarization (Fig. 8, graphs P and Q), an effect culminating at 25% increase during exposure to 30 μM of the drug for the n = 2 readings that could be captured (Fig. 10, graph M). Similar effects could also be inferred at the intermediate (60 bpm) and slow (15 bpm) pacing rates despite the high amount of lacking data, the effects on AP amplitude and depolarization velocity displaying some degree of use-dependency as suggested by the data available at 3 μM (Fig. 10, graphs D through O) while the AP elongating effects globally persisted upon washout. Of note, notwithstanding the few complete data sets, no torsadogenic early after depolarizations (EADs) events were observed in response to R(-)HCQ at any concentration and pacing rate.

Compared to R(-)HCQ, its companion enantiomer S(+)HCQ had qualitatively similar effects on the AP metrics, although quantitatively it affected the AP duration more (notably at the very slow pacing rate) and the resting membrane potential less (see in particular Fig. 10, graphs B and O). The prolongation of the action potential at near complete repolarization was concentration-dependent, reverse use-dependent and only partially reversible upon washout (Fig. 9, graphs P through R, and Fig. 10, graphs M through O). As a quantitative illustration, the median APD90 duration at the concentration of 10 μM was increased by 12% from 220 ms to 246 ms at 180 bpm, by 34% from 358 ms to 481 ms at 60 bpm, and by 56% from 531 ms to 827 ms at 15 bpm. Autogenic activity escaping electrical pacing was seen in two out of six fibers exposed to 30 μM at all rates. Of note, at the extreme slow pacing rate of 15 bpm, abnormal AP exhibiting EADs were seen in one of the six fibers exposed to 10 μM, and in one of the four fibers that could follow pacing in the presence of 30 μM S(+)HCQ, suggesting some degree of torsadogenic potential for this enantiomer beyond 10 μM (see AP waveform in Fig. 9C). Regarding the other AP metrics, S(+)HCQ also substantially decreased the speed of the action potential upstroke depolarization (Fig. 9, graphs J, K and L), an effect that was partially reversible and use-dependent (Fig. 10, graphs G, H and I). At the most rapid pacing rate (180 bpm), a depolarization of the resting potential was observed which culminated at 11% decrease from – 91 mV to – 81 mV in the presence of 30 μM S(+)HCQ (Fig. 9D), a significant and concentration-dependent change, although clearly less pronounced than the effect produced by R(-)HCQ on this same parameter (Fig. 10 A, B and C).

Table 2 compares concentration estimates made by standard LC-MS analytic methods on the R(-)HCQ and S(+)HCQ content in drug-formulated buffer samples collected in the reservoir and at the outlet of the tissue chamber after 20 min perfusion of the Purkinje fibers. The data indicate that none of the tested compounds was lost by retention in the conductive tissue pieces or adherence to components of the experimental setup such as tissue chamber walls or the buffer circulation tubing and pumps.

Table 2.

R(-) and S(+)HCQ content in the formulated buffer samples collected before and after exposure to the Purkinje fibers.

Drug Target (μM) Actual (μM) Loss
R(-)HCQ 0.3 0.29 - 3%
3 2.9 - 3%
10 9.8 - 2%
30 29.6 - 1%
S(+)HCQ 0.3 0.31 +3%
3 3.1 +3%
10 10.1 +1%
30 30.1 <1%
Open in a new tab

4. Discussion

The present experiments indicate that, among the cardiac channels comprising the CiPA panel, the R(-)- and S(+)-enantiomers of HCQ selectively inhibit the human Kir2.1- and hERG-potassium channels, leaving all four other major cardiac currents unaffected up to the highest tested concentrations. We also found that this inhibitory profile can translate into significant alterations of rabbit Purkinje fibers action potential (AP) parameters recorded in an ex vivo situation at concentrations above the therapeutically effective free circulating levels of HCQ.

The automated patch-clamp studies show that the Kir2.1 channels are the most sensitive to inhibition, with IC50 values occurring within a low 1 μM–10 μM concentration range for both enantiomers, while the blockade of hERG channels by R(-)HCQ and S(+)HCQ developed within a tenfold higher concentration range. In contrast, neither R(-)HCQ nor S(+)HCQ affected NaV1.5, CaV1.2, KV4.3 nor KV7.1 channels by a meaningful amount up to 90 μM. This cardiac ion channel inhibitory profile agrees with the profile of chloroquine (CQ), the antiparasitic predecessor of HCQ. Indeed, using a manual patch-clamp approach (Crumb et al., 2016), observed less than 20% inhibition of human recombinant NaV1.5, CaV1.2, KV4.3 and KV7.1 channels exposed to 30 μM CQ, while hERG and Kir2.1 channels where inhibited with IC50 values of 7 μM and 11 μM, respectively. Similarly, in thorough mechanistic study aimed at identifying the biophysical and molecular basis of Kir2.1 blockade by CQ, the IC50 value on the outward component was found to be 9 μM (Rodriguez-Menchaca et al., 2008). Hence, hydroxylation of CQ's side chain, which reduces its toxicity, does not affect its selectivity nor greatly change its potency at those of the CiPA cardiac ion channels that are sensitive to 4-aminoquinolines. More specifically, our data are also fully in line with those recently obtained on a new generation QPatch II platform, with IC50 values for racemic HCQ on hERG and Kir2.1 reaching 17 μM and 30 μM, respectively (Okada et al., 2021). While it is unsure to which extent the differences in IC50 values are significant, the reasons why the channel sensitivity is inverted in both studies is unclear.

Mechanistically, it is tempting to ascribe the non-use-depend depolarizing effects of R(-)HCQ on the Purkinje fibers to its prominent potency at blocking non-voltage-gated Kir2.1 channels. These inwardly rectifying pores are chiefly involved in setting the diastolic resting membrane potential close to the potassium equilibrium potential between consecutive AP firings. Kir2.1 channels also contribute to the repolarization phase of the AP in tandem with hERG and KV7.1 channels (Hibino et al., 2010), providing a way by which R(-)HCQ prolongs the APD90. This prolongation of the AP duration could indeed be observed over the full R(-)HCQ concentration-range at the fastest pacing rate, and up to 10 μM and 3 μM at the intermediate and slow pacing rates, respectively. In the absence of direct inhibition of NaV1.5 channels, the slowing of the AP depolarization velocity and the decrease of the AP amplitude induced by both R(-)HCQ and S(+)HCQ can also be regarded as consequences of the Kir2.1 channel blockade, given that the rising membrane depolarization will engage fast inactivation of a substantially growing fraction of NaV1.5 channels, thereby removing them from the population available in the resting-state ready for activation. This implies that the effects of S(+)HCQ, which is slightly less potent at blocking Kir2.1 channels, on AP amplitude and depolarization velocity would be somewhat less pronounced. This is however difficult to establish given the incomplete sets of measurements we could acquire across pacing rates due to the autogenic activity induced by R(-)HCQ beyond 3 μM.

Overall, we found only a modest degree of enantiomeric separation between R(-)HCQ and S(+)HCQ. Their effects on the cardinal AP metrics were largely similar (see Fig. 10) and in the patch-clamp experiments, their block-potency ratio against Kir2.1 was only two-fold, and hardly reached four-fold against hERG, although in both cases the levorotatory enantiomer was the most potent. The pharmacodynamics of HCQ's enantiomers had previously been compared in other in vitro models without showing potency differences. For example, R(-)HCQ and S(+)HCQ were found equipotent at inhibiting human immune-deficiency virus serotype 1 replication in human T-cell lines and macrophage hybridoma (Chiang et al., 1996). Likewise, the proliferation of Plasmodium falciparum isolates grown in human erythrocytes was inhibited with identical IC50 values by R(-)HCQ and S(+)HCQ (Warhurst et al., 2003). Our present in vitro data also do not suggest that racemic HCQ harbors a clear-cut eutomer and distomer combination of drugs from the arrhythmogenic cardiac safety standpoint. Chirality affects somewhat more meaningfully the pharmacokinetics of HCQ. Indeed, besides the fact that its fate in the body is characterized by very large volumes of distributions and long half-life, the circulating levels at therapeutic doses display enantioselectivity. Specifically, in rheumatoid arthritis patients stably treated with racemic HCQ at the recommended dose, the R(-)/S(+) ratio in blood levels was found to be greater than one in all individuals despite a large two-to three-fold variability between patients (Tett et al., 1994). Moreover, the pharmacokinetics of HCQ were confirmed to be stereoselective in rats (Wei et al., 1995) and rabbits (Ducharme et al., 1994) with no hints for in vivo racemization and a preferential accumulation of the R(-)HCQ enantiomer in ocular tissue (Wainer et al., 1994).

The recent consideration of HCQ as a potential treatment for Covid-19 has generated concerns because of its 4-aminoquinoline structure common to CQ which is known to prolong the QT-segment of the ECG. In a meta-analysis of 28 clinical studies in which patients took HCQ with or without other QT-prolonging drugs, the occurrence of QT-prolongations associated with HCQ monotherapy was below 2%, although patients over 60 appeared significantly more sensitive (Oscanoa et al., 2020). Our in vitro cardiosafety data should be discussed with respect to the circulating drug levels at therapeutic doses. In a population pharmacokinetics model developed from rheumatoid arthritis patients receiving daily doses of 400 mg HCQ, average concentrations in the whole blood reached 2.8 μM (Carmichael et al., 2003). However, recent data comparing whole blood, serum and plasma levels demonstrated that the plasma concentrations of HCQ are two-fold lower than those in blood (Carlsson et al., 2020). Furthermore, in vitro data indicate that the enantiomer of HCQ bind plasma proteins, with the free fraction of R(-)HCQ being about twice as high as that of S(+)HCQ at 63% and 36%, respectively (McLachlan et al., 1993). Given that only the free fraction exercises pharmacodynamic effects such as those described here, a conservative estimate of the therapeutic free plasma concentrations of HCQ enantiomers for comparison with our electrophysiology findings are clearly below 1 μM. Specifically, such a sub-micromolar estimate is comforted by data calculated in a population PK model derived from lupus patients in which mean plasma HCQ ranges between 0.5 μM–0.9 μM for Cmax and 0.3 μM–0.4 μM for Ctrough (Morita et al., 2016). These conservative estimations may however gain relevance in specific circumstances where the cardiac AP repolarization reserve is altered, such as for example in carriers of congenital long-QT syndromes (the prevalence of which could be estimated to be as high as 1:2500 according to recent studies (Schwartz et al., 2009)), or when medications known to prolong the QT interval are administered concomitantly (Roden, 2004).

The present study has limitations, mostly inherent to the in vitro and ex vivo nature of our experimental approaches. First, the small sample size of the Purkinje fibers studies, that was driven by animal ethics considerations and the fastidiousness of manual Purkinje fiber recordings, may reduce the general relevance of our experimental observations. This applies notably to the effects of R(-)HCQ since several parameters could not be properly assessed over all pacing rates and concentrations because of the autogenic activity induced by this enantiomer. It would be relevant to compare the effects of both enantiomers in a more integrated model such as, for example, the rabbit whole-heart preparation perfused according to the Langendorff method, which might offer a better translation to the clinical situation given its higher degree of physiologic integrity (Ellermann et al., 2021). Second, there are uncertainties over the drug effective concentrations when attempting to translate to the clinical situation. Indeed, the defined buffers used in the in vitro patch-clamp and the ex vivo voltage-clamp experiments do not exhaustively mimic the plasma milieu, even if the enantiomers of HCQ bind plasmatic proteins relatively loosely (McLachlan et al., 1993) and are readily soluble, and not particularly suspected of stickiness due to lipophilicity, which can be a source of compound loss in automated patch-clamp experiments (Kramer et al., 2020). Third, the Purkinje fiber studies were performed in standard conditions. However, in real life, HCQ may need to be given to patients with associated pro-arrhythmic risk factors, such as for example hypokalemia, which is well known to promote cardiac arrhythmias (Weiss et al., 2017). It could have been useful to evaluate the effects of R(-)HCQ and S(+)HCQ in the presence of reduced potassium concentrations, as has been recently been done when racemic HCQ was tested in human adult cardiomyocytes. Indeed, the proarrhythmic risk rose significantly when the potassium concentration in the experimental buffer was decreased from 4 mM to 2.5 mM (Jordaan et al., 2021). Finally, the translatability to humans of our Purkinje fibers data needs to consider the fact that I to currents are primarily conducted through KV1.4 channels in rabbits rather than KV4.3 as in humans, a feature that may result in differences in phase 1 repolarization and propensity to engage arrhythmia (Patel and Campbell, 2005).

In conclusion, ion channel inhibitions and cardiac cell membrane alterations were seen with both R(-)HCQ and S(+)HCQ at concentrations in the single to double-digit micromolar range. Given the sub-micromolar free plasma concentrations that can reasonably be estimated for HCQ enantiomers at therapeutic doses, the present in vitro data agree with the low incidence of arrhythmogenic events such as TdP reported over the long-standing clinical practice of HCQ monotherapies in its approved indications. However, caution should be exercised when this drug is given in conjunction with other QT-prolonging drugs or in patients with additional risk factors such as electrolyte imbalance or congenital long QT syndromes.

CRediT authorship contribution statement

Véronique Ballet: designed and directed the studies, interpreted the results and wrote the manuscript.. G. Andrees Bohme: interpreted the results and wrote the manuscript.. Eric Brohan: designed and directed the studies. Rachid Boukaiba: designed and directed the studies. Jean-Marie Chambard: designed and directed the studies. Odile Angouillant-Boniface: generated and curated the data. Thierry Carriot: generated and curated the data. Céline Chantoiseau: generated and curated the data. Sophie Fouconnier: generated and curated the data. Sylvie Houtmann: generated and curated the data. Céline Prévost: generated and curated the data. Brigitte Schombert: generated and curated the data. Laurent Schio: initiated and supervised the project. Michel Partiseti: initiated and supervised the project, interpreted the results and wrote the manuscript, All authors approved the final version and concur with the conclusions.

Declaration of competing interest

All authors are current or former Sanofi employees and may hold shares and/or stock options in the company.

Acknowledgements

The authors are grateful to Dorothée Tamarelle for her expert counseling in the statistical analysis of the data and critical reading of the manuscript. The contributions of Christelle Musso for HPLC quantification of R(-)HCQ and S(+)HCQ containing buffers and of Fiona Ducrey for proofreading of the final manuscript are also gratefully acknowledged.

References

  1. Baczko I., Jost N., Virag L., Bosze Z., Varro A. Rabbit models as tools for preclinical cardiac electrophysiological safety testing: importance of repolarization reserve. Prog. Biophys. Mol. Biol. 2016;121:157–168. doi: 10.1016/j.pbiomolbio.2016.05.002. [DOI] [PubMed] [Google Scholar]
  2. Blaney P.M.B.S.J., Carr G., Ellames G.J., Herbert J.M., Peace J.E., Smith D.I. A practical synthesis of the enantiomers of hydroxychloroquine. Tetrahedron: Asymmetry. 1994;5:1815–1822. [Google Scholar]
  3. Carlsson H., Hjorton K., Abujrais S., Ronnblom L., Akerfeldt T., Kultima K. Measurement of hydroxychloroquine in blood from SLE patients using LC-HRMS-evaluation of whole blood, plasma, and serum as sample matrices. Arthritis Res. Ther. 2020;22:125. doi: 10.1186/s13075-020-02211-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Carmichael S.J., Charles B., Tett S.E. Population pharmacokinetics of hydroxychloroquine in patients with rheumatoid arthritis. Ther. Drug Monit. 2003;25:671–681. doi: 10.1097/00007691-200312000-00005. [DOI] [PubMed] [Google Scholar]
  5. Chiang G., Sassaroli M., Louie M., Chen H., Stecher V.J., Sperber K. Inhibition of HIV-1 replication by hydroxychloroquine: mechanism of action and comparison with zidovudine. Clin. Therapeut. 1996;18:1080–1092. doi: 10.1016/s0149-2918(96)80063-4. [DOI] [PubMed] [Google Scholar]
  6. Crumb W.J., Jr., Vicente J., Johannesen L., Strauss D.G. An evaluation of 30 clinical drugs against the comprehensive in vitro proarrhythmia assay (CiPA) proposed ion channel panel. J. Pharmacol. Toxicol. Methods. 2016;81:251–262. doi: 10.1016/j.vascn.2016.03.009. [DOI] [PubMed] [Google Scholar]
  7. Donovan B.T., Bakshi T., Galbraith S.E., Nixon C.J., Payne L.A., Martens S.F. Utility of frozen cell lines in medium throughput electrophysiology screening of hERG and NaV1.5 blockade. J. Pharmacol. Toxicol. Methods. 2011;64:269–276. doi: 10.1016/j.vascn.2011.09.002. [DOI] [PubMed] [Google Scholar]
  8. Ducharme J., Wainer I.W., Parenteau H.I., Rodman J.H. Stereoselective distribution of hydroxychloroquine in the rabbit following single and multiple oral doses of the racemate and the separate enantiomers. Chirality. 1994;6:337–346. doi: 10.1002/chir.530060418. [DOI] [PubMed] [Google Scholar]
  9. Ellermann C., Wolfes J., Eckardt L., Frommeyer G. Role of the rabbit whole-heart model for electrophysiologic safety pharmacology of non-cardiovascular drugs. Europace. 2021;23:828–836. doi: 10.1093/europace/euaa288. [DOI] [PubMed] [Google Scholar]
  10. Gintant G., Sager P.T., Stockbridge N. Evolution of strategies to improve preclinical cardiac safety testing. Nat. Rev. Drug Discov. 2016;15:457–471. doi: 10.1038/nrd.2015.34. [DOI] [PubMed] [Google Scholar]
  11. Hibino H., Inanobe A., Furutani K., Murakami S., Findlay I., Kurachi Y. Inwardly rectifying potassium channels: their structure, function, and physiological roles. Physiol. Rev. 2010;90:291–366. doi: 10.1152/physrev.00021.2009. [DOI] [PubMed] [Google Scholar]
  12. Kramer J., Himmel H.M., Lindqvist A., Stoelzle-Feix S., Chaudhary K.W., Li D., Bohme G.A., Bridgland-Taylor M., Hebeisen S., Fan J., Renganathan M., Imredy J., Humphries E.S.A., Brinkwirth N., Strassmaier T., Ohtsuki A., Danker T., Vanoye C., Polonchuk L., Fermini B., Pierson J.B., Gintant G. Publisher Correction: cross-site and cross-platform variability of automated patch clamp assessments of drug effects on human cardiac currents in recombinant cells. Sci. Rep. 2020;10 doi: 10.1038/s41598-020-62344-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kuryshev Y.A., Brown A.M., Duzic E., Kirsch G.E. Evaluating state dependence and subtype selectivity of calcium channel modulators in automated electrophysiology assays. Assay Drug Dev. Technol. 2014;12:110–119. doi: 10.1089/adt.2013.552. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Le Marois M., Ballet V., Sanson C., Maizieres M.A., Carriot T., Chantoiseau C., Partiseti M., Bohme G.A. Cannabidiol inhibits multiple cardiac ion channels and shortens ventricular action potential duration in vitro. Eur. J. Pharmacol. 2020;886 doi: 10.1016/j.ejphar.2020.173542. [DOI] [PubMed] [Google Scholar]
  15. Liu H., Yang L., Chen K.H., Sun H.Y., Jin M.W., Xiao G.S., Wang Y., Li G.R. SKF-96365 blocks human ether-a-go-go-related gene potassium channels stably expressed in HEK 293 cells. Pharmacol. Res. 2016;104:61–69. doi: 10.1016/j.phrs.2015.12.012. [DOI] [PubMed] [Google Scholar]
  16. Mazzanti A., Briani M., Kukavica D., Bulian F., Marelli S., Trancuccio A., Monteforte N., Manciulli T., Morini M., Carlucci A., Viggiani G., Cannata F., Negri S., Bloise R., Memmi M., Gambelli P., Carbone A., Molteni M., Bianchini R., Salgarello R., Sozzi S., De Cata P., Fanfulla F., Ceriana P., Locatelli C., Napolitano C., Chiovato L., Tomasi L., Stefanini G.G., Condorelli G., Priori S.G. Association of hydroxychloroquine with QTc interval in patients with COVID-19. Circulation. 2020;142:513–515. doi: 10.1161/CIRCULATIONAHA.120.048476. [DOI] [PubMed] [Google Scholar]
  17. McLachlan A.J., Cutler D.J., Tett S.E. Plasma protein binding of the enantiomers of hydroxychloroquine and metabolites. Eur. J. Clin. Pharmacol. 1993;44:481–484. doi: 10.1007/BF00315548. [DOI] [PubMed] [Google Scholar]
  18. Morita S., Takahashi T., Yoshida Y., Yokota N. Population pharmacokinetics of hydroxychloroquine in Japanese patients with cutaneous or systemic lupus erythematosus. Ther. Drug Monit. 2016;38:259–267. doi: 10.1097/FTD.0000000000000261. [DOI] [PubMed] [Google Scholar]
  19. Okada J.I., Yoshinaga T., Washio T., Sawada K., Sugiura S., Hisada T. Chloroquine and hydroxychloroquine provoke arrhythmias at concentrations higher than those clinically used to treat COVID-19: a simulation study. Clin. Transl. Sci. 2021;14:1092–1100. doi: 10.1111/cts.12976. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Oscanoa T.J., Vidal X., Kanters J.K., Romero-Ortuno R. Frequency of long QT in patients with SARS-CoV-2 infection treated with hydroxychloroquine: a meta-analysis. Int. J. Antimicrob. Agents. 2020 doi: 10.1016/j.ijantimicag.2020.106212. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Patel S.P., Campbell D.L. Transient outward potassium current, 'Ito', phenotypes in the mammalian left ventricle: underlying molecular, cellular and biophysical mechanisms. J. Physiol. 2005;569:7–39. doi: 10.1113/jphysiol.2005.086223. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Jordaan P., Dumotier B., Traebert M., Miller P.E., Ghetti A., Urban L., Abi-Gerges N. Cardiotoxic potential of hydroxychloroquine, chloroquine and azithromycin in adult human primary cardiomyocytes. Toxicol. Sci. 2021;180:356–368. doi: 10.1093/toxsci/kfaa194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Rampe D., Brown A.M. A history of the role of the hERG channel in cardiac risk assessment. J. Pharmacol. Toxicol. Methods. 2013;68:13–22. doi: 10.1016/j.vascn.2013.03.005. [DOI] [PubMed] [Google Scholar]
  24. Roden D.M. Drug-induced prolongation of the QT interval. N. Engl. J. Med. 2004;350:1013–1022. doi: 10.1056/NEJMra032426. [DOI] [PubMed] [Google Scholar]
  25. Rodriguez-Menchaca A.A., Navarro-Polanco R.A., Ferrer-Villada T., Rupp J., Sachse F.B., Tristani-Firouzi M., Sanchez-Chapula J.A. The molecular basis of chloroquine block of the inward rectifier Kir2.1 channel. Proc. Natl. Acad. Sci. U. S. A. 2008;105:1364–1368. doi: 10.1073/pnas.0708153105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Sanguinetti M.C., Jiang C., Curran M.E., Keating M.T. A mechanistic link between an inherited and an acquired cardiac arrhythmia: HERG encodes the IKr potassium channel. Cell. 1995;81:299–307. doi: 10.1016/0092-8674(95)90340-2. [DOI] [PubMed] [Google Scholar]
  27. Schrezenmeier E., Dorner T. Mechanisms of action of hydroxychloroquine and chloroquine: implications for rheumatology. Nat. Rev. Rheumatol. 2020;16:155–166. doi: 10.1038/s41584-020-0372-x. [DOI] [PubMed] [Google Scholar]
  28. Schwartz P.J., Stramba-Badiale M., Crotti L., Pedrazzini M., Besana A., Bosi G., Gabbarini F., Goulene K., Insolia R., Mannarino S., Mosca F., Nespoli L., Rimini A., Rosati E., Salice P., Spazzolini C. Prevalence of the congenital long-QT syndrome. Circulation. 2009;120:1761–1767. doi: 10.1161/CIRCULATIONAHA.109.863209. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Tett S.E., McLachlan A.J., Cutler D.J., Day R.O. Pharmacokinetics and pharmacodynamics of hydroxychloroquine enantiomers in patients with rheumatoid arthritis receiving multiple doses of racemate. Chirality. 1994;6:355–359. doi: 10.1002/chir.530060420. [DOI] [PubMed] [Google Scholar]
  30. Tucker G.T. Chiral switches. Lancet. 2000;355:1085–1087. doi: 10.1016/S0140-6736(00)02047-X. [DOI] [PubMed] [Google Scholar]
  31. Wainer I.W., Chen J.C., Parenteau H., Abdullah S., Ducharme J., Fieger H., Iredale J. Distribution of the enantiomers of hydroxychloroquine and its metabolites in ocular tissues of the rabbit after oral administration of racemic-hydroxychloroquine. Chirality. 1994;6:347–354. doi: 10.1002/chir.530060419. [DOI] [PubMed] [Google Scholar]
  32. Warhurst D.C., Steele J.C., Adagu I.S., Craig J.C., Cullander C. Hydroxychloroquine is much less active than chloroquine against chloroquine-resistant Plasmodium falciparum, in agreement with its physicochemical properties. J. Antimicrob. Chemother. 2003;52:188–193. doi: 10.1093/jac/dkg319. [DOI] [PubMed] [Google Scholar]
  33. Wei Y., Nygard G.A., Ellertson S.L., Khalil S.K. Stereoselective disposition of hydroxychloroquine and its metabolite in rats. Chirality. 1995;7:598–604. doi: 10.1002/chir.530070807. [DOI] [PubMed] [Google Scholar]
  34. Weiss J.N., Qu Z., Shivkumar K. Electrophysiology of hypokalemia and hyperkalemia. Circ Arrhythm Electrophysiol. 2017;10 doi: 10.1161/CIRCEP.116.004667. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Wilson L.J., Mi C., Kraml C.M. A preparative chiral separation of hydroxychloroquine using supercritical fluid chromatography. J. Chromatogr. A. 2020;1634 doi: 10.1016/j.chroma.2020.461661. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from European Journal of Pharmacology are provided here courtesy of Elsevier

RESOURCES